Audity and Unity audio

Audity does not replace Unity’s audio system; it rides it. A playing Sound Graph is an ordinary AudioSource whose content happens to be a compiled graph, and the AudioSource behaves the way it does for any other content. This page states exactly where the seam is: what Unity still does, what the graph does, and the one rule for not doing the same thing twice.

The contract

Audity writes exactly one property on its AudioSource: generator. That slot is also an authoring surface: drop a Sound Graph straight into it and it plays with no Audity components at all - an Audity Source added next to it (Graph field left empty) adopts the slot’s graph and layers its overrides, batching, and quantized starts on top. Everything else on the component is yours, and it works:

AudioSource setting	On Audity output
Volume, Mute	Applies.
Pitch	Applies - the host shifts the whole source, like any clip. Use the Wave Player’s Pitch pin for per-voice offsets (random variation per trigger), which no AudioSource knob can do.
Stereo Pan	Applies (2D).
Spatial Blend, 3D settings (rolloff, min/max distance)	Applies - Unity positions and attenuates the output. See the one rule below.
Output (AudioMixer group)	Applies - routes into the mixer like any source.
Play On Awake	Starts a graph authored on the generator slot (slot-first). With the graph in an Audity Source’s Graph field instead, use Audity Source > Play On Enable.
Loop, AudioClip	Not used - looping and content live in the graph and its Samples.
Doppler Level	Inert on graph output today. The graph’s Doppler node is the Doppler.

The scene’s Audio Listener is authoritative for everything: Unity’s 3D audio uses it directly, and Audity’s Listener Info node reads the same listener.

The one rule: pick your spatializer

You have two legitimate ways to place a sound in space, and they must not be combined on one source:

• Unity 3D: Spatial Blend at 1, a spatial-free graph. Unity (or a spatializer plugin) pans and attenuates the output. Zero graph work; standard behavior. Unity sums every channel to mono before it pans, so author these graphs with a mono Output, which is what a new Sound Graph starts with.
• Graph-authored: Spatial Blend at 0, spatial nodes in the graph (Listener Info → Curve → Gain for falloff, Pan for direction, Doppler for motion). This is the creative option: custom falloff curves, distance driving a filter, authored 7.1 layouts - things no AudioSource knob expresses.

Both at once means the audio is panned and attenuated twice. The Audity Source inspector warns when it sees this combination.

Mixing: use Unity’s mixer

Audity adds no bus system in this release, deliberately - the AudioMixer is the bus tree, and Audity sources are ordinary citizens in it:

• Category volume (the options menu): route sources into mixer groups (Music, SFX, Voice), expose each group’s volume, and drive it with AudioMixer.SetFloat. Nothing Audity-specific required.
• Ducking: the mixer’s Duck Volume effect, sidechained from the dialogue group - or, inside a single graph, the Compressor node’s Sidechain input and Gain Reduction output.
• Snapshots: mixer snapshots work as always (pause states, underwater, menus).
• Code-spawned sounds: AudityPlayback.Play mounts are hidden pooled AudioSources, so pass their routing explicitly: AudityPlayback.Play(graph, position, mixerGroup: sfxGroup). The mount’s physical knobs (volume, 3D blend, rolloff, Doppler) travel the same way, as an AudityMountSpec; see Scripting.

What the graph owns

Everything about how the sound behaves before it leaves the source: content and variation (Samples, Sample Sets), voices and polyphony, per-voice pitch and gain, effects (the graph is the insert chain), musical timing, and the interface your game drives. If a decision is creative, it lives in the graph; if it is routing or scene placement, it lives on the Unity side.